What is it about?
Artificial intelligence is increasingly used to give consumers recommendations, but people do not always respond to AI advice in the same way as human advice. In this article, we examine whether consumers are more persuaded by product recommendations when they come from an AI or a human advisor, and whether the advisor’s confidence and numerical precision matter. Across an online experiment and a field experiment using Meta advertisements, we find that consumers generally preferred recommendations attributed to humans over those attributed to AI. However, recommendations expressed with higher confidence were more persuasive, regardless of whether the source was human or AI. In contrast, presenting confidence with very precise numbers, such as decimals, did not make recommendations more effective. These findings suggest that how AI advice is communicated matters, but that confidence may be more influential than numerical precision in shaping consumer responses.
Featured Image
Photo by Igor Omilaev on Unsplash
Why is it important?
This work is important because AI is increasingly used to guide consumer decisions, yet people may not accept AI advice in the same way as human advice. Our findings show that consumers generally preferred human recommendations, but that confident recommendations were more persuasive regardless of whether they came from a human or an AI. This helps clarify when AI advice is accepted and how it should be communicated. The results are relevant for organizations using AI recommendation systems, especially in digital marketing and online retail contexts.
Perspectives
This article reflects my interest in the human side of AI adoption. As AI systems become more common in marketing and consumer decision-making, it is important to understand not only whether they can provide good recommendations, but also whether people are willing to accept them. I find it particularly interesting that confidence mattered more than numerical precision. This suggests that the way AI advice is communicated may be just as important as the advice itself. I hope this work contributes to a better understanding of how AI-based recommendations can be designed and communicated in ways that people find credible and useful.
Alvaro Chacon
Universidad Tecnica Federico Santa Maria
Read the Original
This page is a summary of: Persuasion in human–artificial intelligence systems: How confidence and precision influence acceptance of recommendations in a consumer context., Technology Mind and Behavior, May 2026, American Psychological Association (APA),
DOI: 10.1037/tmb0000195.
You can read the full text:
Contributors
The following have contributed to this page
